Poster — Fine-Tuning LLMs for Cooperative Tactical Deconfliction of Small UAS
↳ MEIS @ CVPR 2026 Workshop — Multi-Agent Embodied Intelligent Systems Meet the Generative-AI Era · Denver, Colorado, USA
Presented our work on fine-tuning Large Language Models (LoRA / GRPO) for cooperative tactical deconfliction of small UAS, built on a BlueSky simulation-to-language pipeline.
// about this talk
I presented a poster at the 3rd Workshop on Multi-Agent Embodied Intelligent Systems Meet the Generative-AI Era (MEIS @ CVPR 2026), held during CVPR 2026 in Denver, Colorado.
The poster covered our work on fine-tuning Large Language Models for cooperative tactical deconfliction of small Unmanned Aerial Systems — a simulation-to-language data generation pipeline based on the BlueSky air traffic simulator, with parameter-efficient fine-tuning (LoRA and LoRA + GRPO) that improves decision accuracy and reduces near mid-air collisions.
📄 See the corresponding paper on the Publications page, or view the poster.